What is necessary to improve seasonal forecasts and climate change projections of the North American Monsoon?

Size: px

Start display at page:

Download "What is necessary to improve seasonal forecasts and climate change projections of the North American Monsoon?"

Edwin Mark Parker
6 years ago
Views:

1 What is necessary to improve seasonal forecasts and climate change projections of the North American Monsoon? Christopher L. Castro Department of Atmospheric Sciences University of Arizona Presentation to SEACAMS February 18, 2010 Los Mochis, Sinaloa, Mexico. Summer 2004 during NAME. Photo by Peter Rogers

2 Presentation Topics What is the North American Monsoon and why is it important for Mexico and the United States? Can we trust current global models to represent the monsoon? Important factors that influence monsoon variability on climate timescales What is needed to improve monsoon forecasts and projections? Concluding points Acknowlegements: Drs. Dave Gochis, Jae Kyung-Schemm, Tereza Cavazos and other members of the NAME Science Working Group. Also to Dr. Francina Dominguez and Brittany Ciancarelli from UA.

3 What is the North American Monsoon (...in brief) The period of maximum precipitation in northwest Mexico and parts of the southwest United States, including Arizona. Rains begin usually at the end of June in Mexico and several weeks later in July in Arizona and New Mexico. They persist usually until the end of August or beginning of September. Monsoon rainfall is from localized thunderstorms and these can vary a lot in intensity and frequency during the summer. The majority of rainfall is localized in the mountains because the formation of storms depends on the influence of the terrain.

4 Monthly average summer precipitation in northwest Mexico

monthly average precipitation is more than 200 mm in July and

5 The Sierra Madre Mountains during the monsoon The majority of rain is localized in the mountains of the Sierra Madre The monthly average precipitation is more than 200 mm in July and August. Note the rapid increase in precipitation between June and July.

6 Near Topolobampo Farther away from the mountains towards the coast of the Gulf of California, there is less precipitation Also, the precipitation occurs only occasionally and not every day.

7 Importance of the monsoon Example of the state of Sonora, Mexico More than sixty percent of the water in the region comes during the summer. The state depends on summer rainfall for water and agricultural resources.

8 Can we trust current global models to represent the monsoon? What interests us: Seasonal forecasts (from NOAA) Climate Change Projections (from IPCC)

10 IPCC GCM P-E results for the Southwest United States (relative to model climatologies) Seager et al. (2007) My question: How can we trust these results when virtually all these models have little or no representation of summer rainfall?

IPCC models indicate that summer precipitation will not change much.

11 Percentage change in summer precipitaiton at the end of this century by IPCC models (scenario A1B) The level of agreement between these models Current IPCC models indicate that summer precipitation will not change much. But we cannot trust this projection much because of the disagreement among the models.

12 Monthly average precipitation from all IPCC models during the past century Average of historical simulations (sres_20c3m) (Francina Dominguez)

15 The way it could have been Global warming is happening! Buddies For Life

! EXECUSE ME, AL THOSE GCMs AIN T EVEN GOT A

16 The way it is AND THE IPCC MODELS WITHOUT A DOUBT TELL US THAT ALL OF THE SOUTHWEST U.S. WILL TURN INTO A GIANT, BURNING WASTELAND!! EXECUSE ME, AL THOSE GCMs AIN T EVEN GOT A MONSOON IN THEM!! Nemeses

17 So just how s that snake oil, climate changey thingy working out for ya?!

18 To make a climate simulation of the North American Monsoon with an atmoshperic model that has value, it is necessary that the model represent the factors that influence it s variablity good enough. That isn t so easy!

19 Diurnal Cycle of Convection Most important Convective clouds form over the mountains in the morning. By afternoon and everning storms propagate to the west towards the Gulf of California where they can organize into mesoscale convective systems if there is sufficient moisture and instability. It s likely that a resolution less than 5 km is necessary to represent this process correctly in regional models. Global models pretty much fail. (Nesbitt et al. 2008)

20 Intraseasonal variability Includes: Easterly waves Tropical cyclones Low level moisture surges Upper level disturbances Madden Julian Oscillation (Moloney et al. 2008) All these factors can help convection organize and intensify.

21 Can IPCC models represent easterly waves? Their variability from 10 to 20º N during the warm season Observations Lin et al. (2009)

Monsoon Interannual Variability Idea: Atmospheric teleconnections that originate in the western Pacific (and maybe other places) affect the distribution and amount of rainfall, especially in the

22 Monsoon Interannual Variability Idea: Atmospheric teleconnections that originate in the western Pacific (and maybe other places) affect the distribution and amount of rainfall, especially in the early part of the summer. The dominant spatial pattern of precipitation anomalies (SPI) in early summer. Its relationship to large-scale circulation (500-mb height anomalies). Ciancarelli et al. (2009)

23 It is necessary that a model represent this summer teleconnection pattern. If not, there is no hope of making a seasonal forecast that has value. Relationship with atmospheric circulation (500-mb height anomalies).

24 Can we use high resolution atmospheric models to improve forecasts and projections? Some examples

25 Seasonal forecasts of accumulated precipitation during summer 2009 National Center for Environmental Prediction (NCEP) Current global model with low resolution (~200 km) Experimental global model with high resolution (~35 km) (Jae Schemm, NCEP)

26 Low resolution global model Note the increase in precipitaiton in the mountains due to the better representation of the diurnal cycle of convection. High resolution global model

27 Monthly precipitation correlation between NCEP seasonal forecast models and observations in northwest Mexico ( ) High resolution May June July Correlation Low resolution August September October Novermber (Jae Schemm and Bhaskar Jha, NCEP)

28 Regional models are different becasue they require data at their lateral boundaries. We re using NCEP and IPCC models.

29 Value added by a regional atmospheric model Global Model or Atmospheric Reanalysis Grid spacing of Hundreds of km Dynamical downscaling Limited area regional model Grid spacing of Tens of km Value added by this process: Better representation of the land surface and atmospheric circulation Improved representation of atmospheric processes, like convective rainfall

30 What comprises an regional atmospheric model? Precipitation processes Turbulent diffusion Land surface energy balance Radiation Dynamic core Descretized fluid dynamics equations Boundary layer Boundary Conditions (e.g. a GCM)

31 Use of WRF for Downscaling of CFS Reforecasts for Warm Season The version of WRF we use is the Advanced Research WRF (ARW) Model physical parameterizations consistent with those of the existing WRF NWP System at UA. Use NARR soil moisture as an initial condition. Summer reforecasts specifically start at the beginning of April, May, or June of the given year for period WRF simulations start at beginning of May or June and end in August. Only 3 ensemble members available per initialization period, unfortunately! Data from NCEP reanalysis 2 is also being dynamically downscaled to assess the performance of the RCM assuming perfect boundary forcing. The domain for these simulations covers the contiguous U.S. with a grid spacing of 32 km.

32 CFS member Downscaling (TYPE 3) NCEP Reanalysis: Downscaling (TYPE 2) June precipitation solutions for one ensemble member (mm day -1 ) WRF Spectral nudging WRF Lateral boundary nudging only Original Global model

33 1993 Summer precipitation Tucson, AZ Single CFS ensemble member initialized in May WRF downscaled simulation with spectral nudging gives best result! Original CFS model and WRF-CFS downscaled with no interior nudging HAVE NO MONSOON!

34 Climatology of WRF-CFS downscaled simulations with spectral nudging vs. original CFS and observations A better representation of the diurnal cycle of convection explains the dramatic improvement in precipitation by the RCM

35 Climatology of WRF-CFS downscaled simulations with spectral nudging vs. original CFS and observations REGION 2 TUCSON, AZ Southeast Arizona

36 Monsoon Ridge Position at Onset (Late June, July) Climatology delayed Climatology accelerated (Castro et al. 2001)

37 July Precipitation regression coefficient with index based on dominant Pacific SST modes Gridded observed Original CFS Ensemble Average WRF-CFS Downscaled Ensemble Average mm day-1

38 Hot off the press How does it improve using SPI technique as a bias correction? Correlation coef.

39 We are also using WRF in a similar way to make climate change projections Example of WRF-simulated precipitation in July This simulation uses the HadCM3 model as the driving GCM. (Francina Dominguez)

40 We anticipate that all climate simulations with WRF will be completed and analyzed within one or two years. It s very clear that there is a great need for this type of information for decision making with respect to resources in Mexico and the United States.

41 Conclusiones Aunque el monzón afecta la gran área del suroeste del EE.UU. y el noroeste de México, la precipitación es de tormentas que ocurre en la escala local. Modelos actuales globales no representan bien los procesos que influyen la precipitación durante el verano. Estos incluyen el ciclo diurno de convección, variabilidad dentro del monzón, y la variabilidad interanual. Por eso, no debemos confiar en los productos actuales que dependen de modelos globales de baja resolución para representar el monzón norteamericano. Es muy probable que esta conclusión sea la misma en otras regiones durante el verano (e.g. el monzón del America del Sur). Modelos (globales o regionales) de alta resolución mejoran la representación del monzón de tal manera que sus resultados se pueden utilizar para tomar decisiones. Estos productos estarán disponibles dentro de pocos años.

42 Taller de los Impactos del Monzón IPGH aprobó esta propuesta de la Comité Geofísica para un taller en Esta propuesta tienen investigadores de México y los EE.UU. (Drs. Jimmy Adegoke y Tereza Cavazos). También les estoy ayudando a ellos en sus esfuerzos. Solamente $7000 fue aprobado y estos recursos no son suficientes para los gastos del taller. Ya buscamos fondos adicionales de fuentes estadounidenses. Se pidió una extensión del taller hasta Algunos lugares para conducirlo incluyen: En la Cuidad de México o Costa Rica como un taller apartado (abril hasta mayo?) En Brasil durante la reunión del AGU más tarde Le interesa apoyar este taller a un miembro de la Comité Geofísica??

On the Appropriateness of Spectral Nudging in Regional Climate Models

On the Appropriateness of Spectral Nudging in Regional Climate Models Christopher L. Castro Department of Atmospheric Sciences University of Arizona Tucson, Arizona, USA Dynamically Downscaled IPCC model